In- and off-laboratory sources of uncertainty in the use of a serum standard reference material as a means of accuracy control in cholesterol determination

Abstract

 Repeated subsampling or a hierarchical design of experiments combined with an analysis of variance (ANOVA) is demonstrated to be a useful tool in the determination of uncertainty components in amount-of-substance measurements. With the reference material of human serum as investigated here for total cholesterol, besides several in-laboratory sources of uncertainty, a vial-to-vial effect which can be regarded as an off-laboratory source was found to be significant. This knowledge might be essential when the material is used for calibration and for the self-assessment of a laboratory.